DataWorks - Embedding vectorization for data processing is now available.
Sep 20 2025
DataWorksSupports the embedding vectorization of data during the integration process. It simplifies the handling of unstructured data, significantly streamlines the ETL process, and enables efficient knowledge vectorization to support AI scenarios such as RAG.
Content
Target users: all DataWorks users. New feature/specification: Data Integration's single-table offline synchronization now includes an embedding vectorization function. It allows you to extract data from heterogeneous sources like OSS, MaxCompute, and HDFS, transform it into vectors, and load it into vector-capable destinations. These destinations include vector databases such as Milvus, Elasticsearch, and OpenSearch, as well as Hologres vector tables. This capability greatly simplifies the ETL workflow, provides efficient knowledge vectorization, and helps implement AI applications like RAG.
Help Document
https://www.alibabacloud.com/help/dataworks/user-guide/data-integration-supports-embedded-vectorization